Similar to the Rosetta Stone, scholars cracked cuneiform using a monument with the same text in three languages. They first decoded the simpler Old Persian script by identifying repeating patterns like kings' titles, then used that knowledge to unlock the more complex Babylonian text.
Though the scripts look completely different, hieroglyphs emerged after cuneiform was established. This suggests Egyptian travelers or diplomats encountered the *idea* of writing in Mesopotamia. They then developed their own system using culturally relevant pretty pictures instead of adopting abstract cuneiform signs.
Fei-Fei Li's lab believed they were the first to combine ConvNets and LSTMs for image captioning, only to discover through a journalist that a team at Google had developed the same breakthrough concurrently. This highlights the phenomenon of parallel innovation in scientific research.
A recently translated tablet revealed that the Babylonian flood story specified a round boat, like a giant coracle. This design, common on Mesopotamian rivers, is nearly impossible to sink. It prioritizes stability and capacity over navigability, making it the perfect vessel to survive a flood.
Advanced multimodal AI can analyze a photo of a messy, handwritten whiteboard session and produce a structured, coherent summary. It can even identify missing points and provide new insights, transforming unstructured creative output into actionable plans.
Current LLMs abstract language into discrete tokens, losing rich information like font, layout, and spatial arrangement. A "pixel maximalist" view argues that processing visual representations of text (as humans do) is a more lossless, general approach that captures the physical manifestation of language in the world.
While writing enabled literature, its initial large-scale driver was administration. The Sumerians quickly developed sophisticated numeracy and bookkeeping to manage tithes, track goods, and run their city-states. This established accounting as a foundational use of the written word.
An ancient tablet recounting the flood story omits phrases like "he said," which are standard in later texts. This suggests it was a script for a live narrator performing different character voices, capturing literature at the exact moment it was transitioning from oral performance to written text.
Cuneiform began as pictographs for simple records like "three bottles of milk." Its revolutionary leap was using those symbols to represent sounds (syllables), enabling the writing of abstract thought, complex grammar, and literature that pictures alone could not capture.
A cuneiform tablet from 1700 BC, predating the Old Testament by a millennium, tells a nearly identical flood story. The Babylonian version attributes the flood to gods annoyed by human noise, whereas Judean authors later repurposed the narrative to be about a single God punishing humanity for its sins.
Contrary to common perception shaped by their use in language, Transformers are not inherently sequential. Their core architecture operates on sets of tokens, with sequence information only injected via positional embeddings. This makes them powerful for non-sequential data like 3D objects or other unordered collections.